The theory behind fanless systems?

The forum for non-component-related silent pc discussions.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

surfntom
Posts: 12
Joined: Fri Nov 16, 2007 11:34 pm

Post by surfntom » Mon Mar 03, 2008 1:47 am

stmok wrote:Regarding that ionic cooler approach...After reading a bit more, I think I'll avoid that approach until its matured and proven.
I tried it myself, i forgot if i mentioned it earlier. Mine was a failure but i will try it again when i have more time. I believe that the voltage was not high enough for how much air id like to get out of it and i believe the building of the anode and cathode is crucial.

I spoke with some experienced engineers about this topic during interviews and it has been tried in industry but the BIG draw back is the amount of dust that it collects over time.

To answer your previous question regarding FEA and CFD. There are tons of packages out there, i highly recommend NOT using one unless you are familiar with how the methods work. You will hear this a lot "garbage in, garbage out". You can on the other hand get pretty pictures that might look convincing but its not use unless you have a decent grasp as to what is going on. Not to mention these software packages range from 10-100 K. If you are a student or know a student you can get a student version for free most of the time.

you mentioned M-cubed heat pipes. I did use those and they were phenomenal, i would only change the steel brackets that they supply for contact/gripping the pipes. they are quite large for what i wanted and would have liked something else. For the price though you cant really go wrong.

surfntom
Posts: 12
Joined: Fri Nov 16, 2007 11:34 pm

Re: The theory behind fanless systems?

Post by surfntom » Mon Mar 03, 2008 1:58 am

stmok wrote: * Attaching copper block to heatsink? Screws with thermal goop? Epoxy with thermal poop? (The screws approach introduces thermal resistance).
screws should be avoided for heatsinks because you want a good thermal interface. The tightening of screws and even the tapping of threads can (and 99% of the time does) cause a bulge at the surface. This why there is a recommended torque for IGBT mounting screws. Anyway avoid screws, try to use a clamping method like all current cpu heatsinks. If you must screw something together then use a through hole and actually SQUEEZE it somehow.

Most users of TIM use way too much. It sucks at conducting heat but its better than air. With that in mind use as little as possible like i said before.

Good choice of cpu (E8***). Can be had for around 210 online. I'm looking at one myself for an htpc that just recently died (p4 3.0 =/)

You could consider a passive water cooled setup, the benefit of water cooling is that the cpu does not heat up easily due to the water reservoir. It eliminates almost all peaks in temp readings but over time shouldnt be any better than air cooling. You can relocate the heat source to another area though which is nice. This is kind of the same idea as the pc in an aquarium filled with mineral oil. It takes what 4 or 7 hours for the pc to reach peak temp and god knows how long to cool down? Could be your solution depending on how often and how long you use your pc.

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Mon Mar 03, 2008 3:04 am

Since this thread has become filled with unusual ideas, here's another.....

Using a custom-made toilet, and some minor plumbing mods, you could turn the water tank of most toilets into the reservoir for a WC setup. It could be made large enough, with cooling fins, to avoid the need for fans.

Plus.....The water in the system would obviously get changed frequently. And if you wanted to get fancier, the temperature of the water could be lowered if it got too hot, with a thermal-controlled water valve that let in some new, cooler water and flushed the old.

A thermostat in the line could keep the computer coolant at a fixed temp...so it didn't get too cold. If your computer room had a bathroom on the opposite side of the wall...the coolant lines could be very short. It would work.... :lol:

surfntom
Posts: 12
Joined: Fri Nov 16, 2007 11:34 pm

Post by surfntom » Mon Mar 03, 2008 1:18 pm

i dont think many people would like to have the location of their pc dependant on the proximity to the nearest WC ;). interesting idea...

stmok
Posts: 26
Joined: Fri Jul 09, 2004 2:35 am
Location: Sydney, Australia

Post by stmok » Mon Mar 03, 2008 9:32 pm

surfntom wrote:To answer your previous question regarding FEA and CFD. There are tons of packages out there, i highly recommend NOT using one unless you are familiar with how the methods work. You will hear this a lot "garbage in, garbage out". You can on the other hand get pretty pictures that might look convincing but its not use unless you have a decent grasp as to what is going on. Not to mention these software packages range from 10-100 K. If you are a student or know a student you can get a student version for free most of the time.
I had access to these, but I don't see much of a point now, when you can do a few calculations on a spreadsheet. (to get you a good idea of what's going on.)
surfntom wrote:you mentioned M-cubed heat pipes. I did use those and they were phenomenal, i would only change the steel brackets that they supply for contact/gripping the pipes. they are quite large for what i wanted and would have liked something else. For the price though you cant really go wrong.
Awesome, thanks for the feedback on that.
surfntom wrote:screws should be avoided for heatsinks because you want a good thermal interface. The tightening of screws and even the tapping of threads can (and 99% of the time does) cause a bulge at the surface. This why there is a recommended torque for IGBT mounting screws. Anyway avoid screws, try to use a clamping method like all current cpu heatsinks. If you must screw something together then use a through hole and actually SQUEEZE it somehow.

Most users of TIM use way too much. It sucks at conducting heat but its better than air. With that in mind use as little as possible like i said before.
OK...Maybe a combination of clamping and a tiny bit of TIM is sufficient?

I was originally thinking, when you attach the copper blocks to the heatsink, you use epoxy on the edges of the blocks and tiny bit of TIM for the inner area. Clamp it down while it sets.
Bluefront wrote:Since this thread has become filled with unusual ideas, here's another.....
No water cooling thanks...LOL, toilet. :lol:

I prefer if we stick to something highly reliable. ie: no moving parts as much as possible. Hence the reason why I'm leaning towards heatsinks and heatpipes.

I have a few questions though...

Does the length of the heatpipe drastically affect the thermal resistance?

What about the shape that its bent to?

I read somewhere about some flexible heatpipe solutions, but I couldn't find who sells these. Anyone know?

The challenge in heatpipe/heatsink approach seems to be getting that heat out of the case to the heatsink, but still be able to access the PC hardware inside. ie: Those removable case panels.

...Something inside me says I should just give up and buy a pre-made silent solution like Tranquil PC, mCubed, etc. Maybe use that refurbished ThinkPad I recently bought...Grab a port replicator for it, and be done with it....I don't know.

surfntom
Posts: 12
Joined: Fri Nov 16, 2007 11:34 pm

Post by surfntom » Mon Mar 03, 2008 11:48 pm

stmok wrote: Does the length of the heatpipe drastically affect the thermal resistance?

What about the shape that its bent to?

I read somewhere about some flexible heatpipe solutions, but I couldn't find who sells these. Anyone know?
the heatpipes are super duper light so you can bend them even without trying. they have thin copper walls so you can imagine how easily they bend. I would say they are much lighter than you would expect. I bent mine by hand making sure not to put a kink in it to 90 degrees easily and had no problem. As long as you make sure not to put a kink you will be fine.

The length shouldnt be much of a concern considering they are not ordered by the foot so you only have a few choices and neither of them are long.

So the answer to you flexible heatpipes is that the m-cubed heat pipes are flexible unless there is some new ultra flexible ones elsewhere.

You also mentioned epoxy on the heatsinks, i dont know how that works out but it is not very good if you ever need to replace it. A thin layer of tim and some good solid pressure should do just fine. Lapping is recommended.

I agree with you that some excel hand calcs definitely gets you in the ball park but if you start to make things the slightest bit more in depth then you realize that your excel sheet grows very fast. I personally use the software to visualize how air will flow given certain geometries rather than trying to get actual temps since the heat transfer devices (heatsinks) are off the shelf and you cant really modify them with ease.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Tue Mar 04, 2008 1:33 am

...Something inside me says I should just give up and buy a pre-made silent solution like Tranquil PC, mCubed, etc. Maybe use that refurbished ThinkPad I recently bought...Grab a port replicator for it, and be done with it....I don't know.
it really depends how much computing and graphics power you need. if you can make do with a Sempron/Celeron and onboard graphics, something like this:

http://www.silenthardware.de/forum/inde ... opic=23348

is trivially easy to put together. however if you need quad-core and game-ready graphics, then it is a problem.

BillTodd
Posts: 132
Joined: Wed Apr 05, 2006 4:10 am
Location: Colchester (nuked in 1984)
Contact:

Post by BillTodd » Tue Mar 04, 2008 7:20 am

Does the length of the heatpipe drastically affect the thermal resistance?
The thermal resistance is caused by the internal wick construction:
i.e. high to low: Mesh, Sintered, Grooved.

Think about the path that the heat takes from surface of the pipe to the working fluid. Grooved walls are going to present less resistance to heat than sintered spheres or a wire mesh.
What about the shape that its bent to?
Bending the pipe affects the wick, if the wick fails the pipe will not work. So, fewer bends are better.

Make sure you follow the manufacturers minimum bend radius spec. Use a proper pipe bender. The internal fluid (usually water) is at a low pressure and the pipe walls are thin, making them very easy to kink them.

The important thing to remember is that there is fluid flow inside the pipe, so anything that reduces the liquid flow (from cool end to hot end) will effect how well the pipe works. Some mesh/sintered wicks will work against gravity (i.e. up hill), but all will work better with gravity (i.e. cold end above the hot)

spookmineer
Patron of SPCR
Posts: 749
Joined: Sat Nov 11, 2006 6:02 pm

Post by spookmineer » Tue Mar 04, 2008 2:09 pm

I stumbled accross this article in another forum:

100% passive case guide. A good read.

Used components (a bit old, project from summer 2007):
* Opteron 170 @ 2,4 GHz (TDP 89 W)
* GeForce 6800GT
* Asus A8N-VM CSM
* Kingston 400 MHz DDR ECC
* Seagate Barracuda 320 GB

In retrospect, the writer thinks the mainboard was faulty and the original lay out would have worked fine.

BillTodd
Posts: 132
Joined: Wed Apr 05, 2006 4:10 am
Location: Colchester (nuked in 1984)
Contact:

Post by BillTodd » Tue Mar 04, 2008 3:24 pm

Opteron 170 @ 2,4 GHz (TDP 89 W)
:shock: I've been struggling* with a 59.3w P4, 89 watt - no wonder he was having problems :lol:


(* I think I've cracked it - 44'C [+22'C] ATM @ full power on the test bench 8) )

EndoSteel
Posts: 308
Joined: Sun Jan 25, 2004 7:50 am
Location: Moscow, Russia
Contact:

Post by EndoSteel » Wed Mar 05, 2008 9:52 am

stmok
Never heard of flexible heatpipes but flexible _thermosyphons_ were once mentioned at overclockers.com. I don't know how did they fabricate them. The problem with flexibility is you need very special pipes to achieve it. A pipe has to be soft in order to bend easily and it has to be vacuumed in order to provide phase-change heat transport at reasonably low temps. Try to pump air out of a rubber hose and you'll understand what I mean. It will collapse. Even a stainless steel bellow tends to compress and straighten under vacuum so, as mentioned, I don't know how did they do it. [/b]

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Wed Mar 05, 2008 3:14 pm

High-pressure rubber hoses have been used in automotive AC systems for years. They are subjected to vacuum on the low-pressure (suction) side, and pressure on the high side.....all without bulging or collapsing. Flexible heat-pipes using a rubber joint are a real possibility.....

Post Reply